The scales of standard barometers are usually divided into half-tenths, or .05, of an inch, as represented, in fig. 5, by AB. The vernier, CD, is made equal in length to twenty-four of these divisions, and divided into twenty-five equal parts; consequently one space on the scale is larger than one on the vernier by the twenty-fifth part of .05, which is .002 inch, so that such a vernier shows differences of .002 inch. The vernier of the figure reading upwards, the lower edge, D, will denote the top of the barometer column; and is the zero of the vernier scale. In fig. 4, the zero being in line exactly with 29 inches and five-tenths of the fixed scale, the barometer reading would be 29.500 inches. It will be seen that the vernier line, _a_, falls short of a division of the scale by, as we have explained, .002 inch; _b_, by .004; _c_, by .006; _d_, by .008; and the next line by one hundredth. If, then, the vernier be moved so as to make _a_ coincide with _z_, on the scale, it will have moved through .002 inch; and if 1 on the vernier be moved into line with _y_ on the scale, the space measured will be .010. Hence, the figures 1, 2, 3, 4, 5 on the vernier measure hundredths, and the intermediate lines even thousandths of an inch. In fig. 5, the zero of the vernier is intermediate 29.65 and 29.70 on the scale. Passing the eye up the vernier and scale, the second line above 3 is perceived to lie evenly with a line of the scale. This gives .03 and .004 to add to 29.65, so that the actual reading is 29.684 inches. It may happen that no line on the vernier _accurately_ lies in the same straight line with one on the scale; in such a case a doubt will arise as to the selection of one from two equally coincident, and the intermediate thousandth of an inch should be taken. For the ordinary purposes of the barometer as a "weather-glass," such minute measurement is not required. Hence, in household and marine barometers the scale need only be divided to tenths, and the vernier constructed to measure hundredths of an inch. This is done by making the vernier either 9 or 11-10ths of an inch long, and dividing it into ten equal parts. The lines above the zero line are then numbered from 1 to 10; sometimes the alternate divisions only are numbered, the intermediate numbers being very readily inferred. Hence, if the first line of the vernier agrees with one on the scale, the next must be out one-tenth of a tenth, or .01 of an inch from agreement with the next _scale_ line; the following vernier line must be .02 out, and so on. Consequently, when the vernier is set to the mercurial column, the difference shown by the vernier from the tenth on the scale is the hundredths to be added to the inches and tenths of the scale.